Australia accuses OpenAI of having accessed health data via AI agents

The artificial intelligence (AI)-based agent accessed public and non-public files on the website that listed Medicare statistics, Albanese told reporters in New York on Wednesday. The agent found a way to circumvent the data protection precautions "and, if you will, didn't take 'no' for an answer."

The situation is “unacceptable,” said Albanese. He spoke to OpenAI boss Sam Altman "to express Australia's deep concern about this incident." He also expressed his “disappointment that the company took far too long to inform the government about what was happening.”

"It took until September 10th for any notification to be made," Albanese said. This was an email that was “merely sent to the general mailbox.”

According to OpenAI, “several Australian government websites” were affected

OpenAI announced a short time later that, according to a recent review, “several Australian government websites” had been affected. “Activities related to various Australian government websites and services” were identified.

During an internal assessment, OpenAI models "attempted to retrieve answers and available statistics to questions about Australia," the statement continued. "Our models carried out actions that we had not intended." The information accessed included health statistics and internal file names. There is no evidence that patient files were accessed.

Anthropic boss Dario Amodei and Altman recently called for “slowing down the pace at which we improve the capabilities of AI models.” The “time gained” should be used sensibly to better understand and contain the risks arising from AI. The reason was several incidents in which AI systems broke out of their test environments and attacked other companies or platforms via the Internet. This fueled fears of potentially uncontrollable AI models.

At the same time, however, there is fierce competition in the AI ​​industry - even with inexpensive models from China. US companies are also under increasing pressure to turn their gigantic investments in AI infrastructure into profits.

This fall may decide whether we will permanently control AI supermachines. Even company bosses are calling for more rules. But Donald Trump and Xi Jinping pursue their own goals. You can find out more about this here in the current SPIEGEL cover story.